Time-Varying Evolution and Impact Analysis of Forest Tourism Carbon Emissions and Forest Park Carbon Sinks in China

Abstract

Forests are an important part of natural resources and play an important role in carbon sinks. We measured carbon sinks in provincial forest parks using data from four forest inventory surveys in China and the forest stock expansion method. Carbon emissions from forest tourism were also estimated using energy statistics and forest park tourism data. On this basis, spatial analysis was used to summarize the spatial and temporal evolution of the carbon balance and the analysis of influencing factors. The results show the following: (1) With the passage of time, the carbon emissions from forest tourism in all provinces have increased to different degrees, and the national forest tourism carbon emissions have increased from 1,071,390.231 (million tons) in 2003 to 286,255,829.7 (million tons) in 2018; spatially, the distribution of carbon emissions from forest tourism is uneven, with an overall high in the south and low in the north, and a high in the east and a low in the west. (2) The carbon sink of forest parks showed a trend of gradual growth and spatially formed a spatial pattern of high in the northeast and low in the southwest, which is consistent with the distribution of forest resources in China. (3) For forest tourism carbon emissions, the total number of tourists, tourism income, and playing roads are significant influencing factors, and the baseline regression coefficients are 0.595, 0.433, and 0.799, respectively, while for forest park carbon sinks, the number of forest park employees can play a certain positive role in carbon sinks, with the regression coefficient being 1.533.

1. Introduction

Against the backdrop of global climate change, controlling carbon emissions and enhancing the capacity of carbon sinks have become a core issue of common concern to all countries [1,2]. Global climate change has not only brought about a series of environmental problems such as extreme weather and sea-level rise but has also had a profound impact on the economic development and quality of life of human society [3,4]. In response to global climate change, countries have adopted a series of strategies focused on reducing carbon dioxide emissions, including joint efforts to reduce greenhouse gas emissions through the Global Compact [5]. As the world’s largest carbon-emitting economy, the Chinese government has made a commitment to the world. It pledged to peak carbon emissions by 2030 and to become carbon neutral by 2060 [6].

The introduction of the “dual-carbon” goal signals the Chinese Government’s determination to ameliorate climate change. However, the realization of this goal requires sustained promotion of greenhouse gas emission reduction in terms of both reducing “carbon sources” and increasing “carbon sinks”. Reducing “carbon sources” mainly means reducing the use of fossil fuels and controlling industrial emissions, as well as gradually developing a low-carbon and green economic growth model. On the other hand, increasing “carbon sinks” includes increasing the carbon absorption capacity of ecosystems such as forests, grasslands, and wetlands [7]. In terms of reducing carbon emissions, the contribution of tourism to carbon reduction has been increasingly emphasized, as tourism is not only an important part of the global economy, but was also once strongly promoted as a clean industry [8,9,10,11]. Tourism has been called a “smokeless industry” mainly because it pollutes less and has less direct impact on the environment than traditional industries and manufacturing industries that are high in pollution and emissions. However, although tourism does not directly engage in industrial production and avoids emissions of sulfur dioxide and nitrogen oxides, its impact on carbon emissions cannot be ignored. Numerous studies have shown that the share of carbon emissions from the tourism industry in carbon emissions has been growing, and the pressure on the environment has been increasing [12,13]. According to the statistics of the United Nations World Tourism Organization (UNWTO), the global tourism industry accounts for about 8% of the world’s total carbon emissions every year. Tourists’ transportation, accommodation, catering, and other activities generate a large amount of carbon dioxide emissions and have a negative impact on the environment, which should attract our attention. In China, more research on tourism carbon emissions has focused on the tourism sector as a whole, such as summarizing spatial and temporal patterns by accounting for differences in provincial tourism carbon emissions, which have been shown to generate more carbon emissions in regions where tourism markets are booming [14]. These carbon emissions are caused by the sharp rise in the number of tourists, especially the energy consumption triggered by the transportation sector being the most intuitive cause [15]. Interestingly, this phenomenon is also validated in the international tourism industry, where inbound tourism and transportation specifically reinforce the carbon emissions of tourism, especially in those developed countries [16]. It goes without saying, therefore, that we believe that by promoting eco-tourism and green tourism, we can reduce the negative impact of tourism activities on the environment.

Of course, achieving carbon neutrality requires increasing carbon sinks in addition to reducing carbon emissions. In terms of increasing carbon sinks, promoting the development of forest carbon sinks is recognized as one of the important ways to mitigate global climate change [17,18]. As the “green lungs” of the earth, forests maintain the carbon balance in the atmosphere by converting carbon dioxide into oxygen and organic matter through photosynthesis. In addition, forest soils store a large amount of organic carbon, making them an important carbon storehouse. Therefore, forest parks have an irreplaceable role in combating climate change [19]. According to the Global Forest Resources Assessment report published by the Food and Agriculture Organization of the United Nations (FAO) in 2020, the global forest carbon stock accounts for about 77% of the global vegetation carbon stock [20]. This finding shows that forests occupy a dominant position in global carbon storage. Enhancing the carbon storage capacity of forests not only effectively reduces the concentration of carbon dioxide in the atmosphere but also helps to address global climate change [21]. In addition, the development of forest carbon sinks not only contributes to the realization of climate goals but also brings multiple ecological benefits [22]. For example, forests can provide biodiversity habitats, protect water sources, prevent soil erosion, and support the livelihoods of many communities. At the same time, forests provide very important sites for ecotourism, and in the process have evolved a specific mode of tourism known as forest tourism. Specific forest areas with good ecosystems and value for tourism development are also defined as forest parks [23,24,25]. In addition to their most basic recreational value, China’s forest parks have received protection and restoration from the Chinese government and often have a certain landscape character that can provide some artistic value [26]. The Chinese government has made great efforts to protect forest resources and develop forest tourism, which has become a favored mode of travel for most tourists. As a result, the topic of forest parks has also gradually entered the field of vision of scholars: Some studies have discussed the factors affecting the satisfaction of forest tourism by designing tourist questionnaires, collecting research data, and discussing them through structural equation modeling and have found that a good ecological environment and rich forest resources can attract tourists [27]. Further, researchers have begun to focus on the relationship between place attachment and environmentally responsible behavior in forest tourism, which suggests that booming forest tourism has also brought some pressure on the local ecological environment [28]. Forest parks have also become popular destinations visited by a large number of tourists, and the carbon emissions generated by tourist activities put significant pressure on the carbon sink function of forest parks [29,30,31].

Therefore, it is now certain through the above literature that the development of tourism puts some pressure on the ecological carrying capacity of forests (especially carbon sinks). However, the development of forest tourism is again a form of low-carbon tourism, as it replaces the excessive carbon emissions of traditional tourism. But what exactly is the relationship between forest tourism carbon emissions and forest park carbon sinks in the synergistic development of forest tourism development and forest park protection in various regions (in this case, different provinces in China)? Are carbon sinks greater than carbon emissions or are carbon emissions greater than carbon sinks? In other words, what the status is of this carbon balance is worth discussing. Second, in addition to accounting for these elements at the provincial level, given the differences in the distribution of forest resources and provincial characteristics in China, what kind of spatial distribution characteristics exist? Are they consistent with traditional perceptions, and what is the clustering pattern? And what are these drivers? What is the direction of evolution over time? These are the research objectives and research questions of this study. Therefore, in the process of the carbon neutrality target, it is crucial to know how to balance the pressure of ecotourism on forest parks under the dual regulation of carbon emission reduction and carbon sink enhancement; in other words, we hope to explore the dynamic spatial relationship between forest tourism carbon emissions and forest park carbon sinks, in order to formulate scientific and reasonable low-carbon tourism and then to realize the basic balance of carbon sinks and carbon emissions of forest tourism in the region, as well as to ensure that the development of ecotourism will not be carried out without sacrificing the basic function of forest carbon sinks [32,33,34].

The marginal contributions of this paper may be as follows: Using forest inventory data, the carbon sink of forest parks at the provincial level was assessed for the first time, and its spatial and temporal evolution characteristics were summarized. The forest tourism carbon emission data of each province were accounted for and the pressure they caused on forest park carbon sinks was examined. An exploratory geographical analysis of the spatial differentiation of forest park carbon sinks, forest park carbon emissions, and forest tourism carbon balance was conducted.

2. Materials and Methods

2.1. Overview of the Study Area

In this paper, we chose the 31 provinces in mainland China as the study area for many reasons: first, considering the scarcity of forest data, our research design often relies on existing data, and the “National Forest Inventory Data” is the closest to the real situation of forests. Therefore, we analyzed China’s forest park carbon sinks and forest tourism carbon emissions on a provincial level. Of course, although the analysis was carried out at the provincial level, it was also necessary to visualize each forest park (because it involves the number of forest parks, and often there are more forest parks in these forest-rich areas). To this end, we drew Figure 1, which shows the distribution of the number of forest parks in China in 2003 and 2018, respectively. Of course, the list of these forest parks comes from information published annually by the State Forestry Administration. Correspondingly, some of the most relevant objective factors involving carbon emissions from tourism should be shown. Here, we chose the result of the total number of visitors to forest parks, and Figure 2 reports the total number of visitors to forest parks in each province in 2003 and 2018. These data are from the “China Forestry and Grassland Statistical Yearbook (2003/2018)”.

It can be seen that both the development of forest tourism and the number of forest parks have shown tremendous growth over the 15-year period 2003–2018. Moreover, the number of tourists and tourism income have increased over time, which undoubtedly indicates that forest tourism is becoming a new type of tourism. Likewise, the number of forest parks has grown dramatically, with more forests being established as forest parks, which are protected and also used to provide recreation.

2.2. Data Sources

With regard to forest tourism data, the data on the number, area, number of tourists, geospatial location, and forest type of national forest parks were obtained from the China Forestry and Grassland Statistical Yearbook, the National Forest Resources Inventory Report and the China Forestry Network (http://www.Forestry.gov.cn/ accessed on 22 August 2024) (2003, 2008, 2013, 2018) Data on forest area and volume at the provincial level are from the China Environmental Statistics Yearbook. The data on forest carbon sinks were selected from six forest inventories in China in 2003, 2008, 2013, and 2018, and the data were obtained from the National Forest Inventory Report and the China Forestry Statistical Yearbook. Data were obtained from the National Forest Resources Inventory Report and the China Forestry Statistical Yearbook. Since it is provincial panel data, there are not many missing values, and the only ones that are missing are filled in using interpolation (which may produce small errors in the measurements).

2.3. Measuring Carbon Emissions from Forest Tourism

For the accounting of carbon emissions from tourism in forest parks, we used the product of the number of visitors to forest parks and the carbon intensity of tourism per capita to measure. Here, the most important element is to determine the calculation of tourism carbon emissions, and we chose the most common “top-down” method [35,36]. Because of the unavailability of formal data on the greenhouse gas emissions from China’s tourism sector, it is crucial to use the tourism coefficient for accurate estimations. Referring to [35], we first cataloged the emissions associated with energy consumption in the tourism industry, and then calculated their equivalents in terms of standard coal emissions [35]. This result was then adjusted with a tourism factor (TF), which represents the ratio of provincial gross tourism revenue (TR) to provincial gross domestic product (GDP):

$$TF={\displaystyle \frac{TR}{GDP}}$$

(1)

Our primary data source is the China Energy Statistics Yearbook. Tourism is the main economic sector for businesses, such as long-distance and local transportation, accommodation, food and beverage, entertainment, shopping, and telecommunications. These sectors are aligned in the Yearbook with categories such as “transportation, storage, and postal services” and “wholesale and retail trade, accommodation, and food services”. In order to record expenditures that are not directly observed, we also included the category “other” in our statistical analysis. The Yearbook provides a broad categorization of fossil fuel consumption in the major categories of crude oil, coal, and natural gas, which are further broken down into subcategories such as raw coal, coking coal, and washed coal. However, certain fossil fuels, particularly lubricants, are used negligibly in the tourism industry, accounting for less than 1% of total oil consumption. The data also have many missing values, which challenged the completeness of our panel data. Simple interpolation methods are not sufficient to address these deficiencies and may increase the risk of experimental inaccuracies. Ultimately, our study considered the consumption of five primary energy sources: raw coal, gasoline, diesel, natural gas, and electricity.

Table 1. Conversion factor of standard coal.

Energy Type	Conversion Factor	Unit
Raw Coal	1.4286	kgce/kg
Gasoline	1.4714	kgce/kg
Diesel	1.4571	kgce/kg
Natural Gas	1.3300	kgce/m3
Electric Power	0.1229	kgce/(kW·h)

Note: These are the official data published by China’s Ministry of Industry and Information Technology (reference link: http://zsjgj.zhoushan.gov.cn/art/2022/6/10/art_1496911_58920733.html accessed on 22 August 2024).

reports the standard coal conversion factors (CF) for these five main energy categories.

Using these data and with the help of the standard coal conversion coefficients in Table 1, we can estimate the total energy consumption of the tourism-related industries and convert the energy demand of the tourism industry into carbon emissions by using the standard coal equivalent coefficient of energy and the carbon emission coefficient (k = 2.45), which is calculated by the formula below.

First, the associated total energy consumption (EC) was calculated by converting and summing the type j energy use CF_j used by the type i industry to standard coal consumption. Further, the total energy consumption of the tourism sector, i.e., TEC, was obtained by multiplying EC with the TF. Finally, the tourism carbon emissions (TCE) were obtained by means of the standard coal emission factor k:

$$EC={\displaystyle \sum _{i=1,j=1}^{n,m}{E}_{ij}\times C{F}_j}$$

(2)

$$TCE=k(EC\times TF)$$

(3)

By now, we have obtained the TCE of tourism carbon emissions for each province. In order to calculate the forest tourism carbon emissions (FTCE), we need to calculate the per capita tourism carbon intensity (TCI) and the number of tourists in forest parks (NTFP). In this case, the TCI is expressed through the ratio of the TCE to the total number of tourists (NT) in the year, which is derived from the Tourism Statistics Yearbook. The NTFP data come from the Statistical Reporting System for the Annual Construction and Operation of Forest Parks:

$$\{\begin{array}{l}FTCE=TCI\times NTFP\\ TCI=TCE/NT\end{array}$$

(4)

2.4. Calculation of Carbon Sequestration in Forest Parks in Various Provinces

In this paper, forest carbon sinks are accounted for by the forest stock expansion method [37,38]. In the formula, C_S, C_V, C_B, and C_F are the carbon sinks of different forest structures in order, which are forest soil carbon sink, understory vegetation carbon sink, forest biomass carbon sink, and overall forest carbon sink, respectively. S_ij, C_ij, V_ij represent the forest area, forest carbon density, and forest stock per unit area of forest type j in category i area, respectively, with the following equations:

$$C_F=C_S+C_B+C_V$$

(5)

$$C_F={\displaystyle \sum (S_{ij}\times C_{ij})}+\alpha {\displaystyle \sum (S_{ij}\times C_{ij})+\beta }{\displaystyle \sum (S_{ij}\times C_{ij})}$$

(6)

$$C_{ij}=V_{ij}\times \sigma \times \rho \times \gamma$$

(7)

where α, β, σ, ρ, and γ represent the understory plant carbon conversion factor, forest carbon conversion factor, microbial content expansion factor, volume factor, and carbon content, respectively. The conversion factors are based on the United Nations Intergovernmental Panel on Climate Change default parameter values. The values for α and β were 0.195 and 1.244, respectively. σ was calculated as 1.90, ρ as 0.50 t/m³, and γ as 0.50. After obtaining the forest carbon sequestration of each province, we further calculated the carbon sequestration of forest parks in each province based on the ratio of forest park area to forest area, recorded as forest park carbon sink. On this basis, the carbon sink of forest parks is measured according to the ratio of forest area to forest area:

$$C_{FP}=(are{a}_{FP}/are{a}_F)\times C_F$$

(8)

where C_FP is the forest park carbon sink, area_FP is the forest park area, and area_F is the forest area. Further, we obtained the forest tourism carbon balance (FTCB) by subtracting forest tourism carbon emissions from the calculated forest park carbon sinks.

2.5. Spatial Correlation Analysis

Exploratory spatial data analysis (ESDA) mainly uses geovisualization techniques to reveal the characteristics of spatial data. It is often used to identify spatial data distribution patterns, aggregation hotspots, and spatial heterogeneity. Spatial autocorrelation methods are categorized into global spatial autocorrelation (Global Moran’s I) and local spatial autocorrelation (Local Moran’s I). Global Moran’s I detects spatial clustering or outliers and finds differences and correlations in the spatial distribution. The value is (−1, +1), and when the global Moran’s I index is >0, the larger the value, the more significant is the spatial positive correlation; when the global Moran’s I index is <0, the smaller the value, the more significant is the spatial variability. Global spatial autocorrelation analysis uses only one value to reflect the average degree of spatial variation between study regions, but it cannot describe in detail the specific spatial correlation patterns between objects within a region and does not take into account spatial heterogeneity; thus, it cannot reflect the local spatial correlation within a geographic unit. Therefore, it is necessary to use local autocorrelation to determine specific clustering situations and to describe the correlation of attribute values in adjacent spaces by performing local autocorrelation analysis [39,40]. The corresponding equations are as follows:

$$I_G={\displaystyle \frac{n{\displaystyle \sum _{i=1}^n{\displaystyle \sum _{j=1}^{n}w{i}_j(y_i-y^{\prime })(y_j-y^{\prime })}}}{({\displaystyle \sum _{i=1}^n{\displaystyle \sum _{j=1}^{n}w{i}_j)}}{\displaystyle \sum _{i=1}^n{(y_i-y^{\prime })}^2}}}$$

(9)

$$I_L={\displaystyle \frac{(y_i-y^{\prime }){\displaystyle \sum _{i=1}^n{w}_{ij}(y_j-y^{\prime })}}{{\displaystyle \sum _{i=1}^n{(y_j-y^{\prime })}^2/n}}}$$

(10)

where I_G, I_L, and n are the global and local Moran’s I indices and the total number of regions, respectively, y′ is the mean value of the corresponding index for the region, y_i and y_j denote the observed values of the corresponding attribute for the i and j region, respectively, and w_ij is the element of the spatial weighting matrix describing the correlation of the spatial objects between the i_th and j_th points. ESDA and Moran’s index are common methods widely used in spatial evolution analysis. Here, we applied these two methods to the spatial pattern induction and evolutionary analysis of CFP and FTCE. Among them, the global Moran index can well describe whether there is a correlation between forest tourism carbon emissions and forest park carbon sinks in these provinces, which is important because it is the basis and prerequisite for conducting pattern analysis. As mentioned above, if the coefficient remains consistently positive, it indicates the existence of a positive correlation, and for a single province, what is the clustering relationship between its own CFP and FTCE and the surrounding neighboring areas? If a region has carbon sinks or emissions that are similar to those of its surrounding areas and significantly higher than those of other regions, it indicates the formation of high–high clusters here. This indicates that there may be high carbon sinks and emissions in this region, and a clustering effect has formed spatially.

2.6. Econometric Methodology

After accounting for carbon sinks in forest parks and carbon emissions from forest tourism, we may be able to extrapolate the factors that influence these results from inherent observations and basic knowledge; e.g., forest carbon sinks are higher in more forested areas, but there are other factors that may be potentially unobserved. These influences on carbon sinks and emissions may be subtle but real. In general, regression analysis is taken to examine the differences in the impact of different factors on carbon sinks and emissions, and here we adopted an econometric model using a two-way stationary model and a variety of tests to ensure that the regression results are robust, i.e., that the results are non-random. Specifically, we need to set up two two-way stationary models to examine the influences of carbon emissions and carbon sinks, respectively.

The first is the analysis of factors affecting carbon emissions from forest tourism:

$$FTC{E}_{it}=\alpha +{\beta }_1{x}_1\cdots +{\beta }_3{x}_3+v_t+u_i+\epsilon$$

(11)

Second, we have to analyze the influencing factors affecting the carbon sink of the forest park, and the set equations are as follows:

$$CF{P}_{it}=\gamma +{\lambda }_1{z}_1\cdots +{\lambda }_3{z}_3+v_t+u_i+\epsilon$$

(12)

Here, β and γ are the regression coefficients of the respective equations, while x₁ to x₃ represent the three influences on forest tourism carbon emissions, similarly, z₁ to z₃ represent the three influences on carbon sinks in forest parks, and lastly, v and u denote the time effect fixed and individual effect fixed, respectively. Specifically, we referred to relevant studies and selected the following variables as potential influencing factors, as shown in

Table 2. Analysis of influencing factors.

Variable	Characterization of Variables	Unit
x1	Total number of visitors to forest parks	10,000 people
x2	Total revenue of forest parks	billions
x3	Total mileage of the tour roads	kilometers
z1	Number of forest park employees	1 person
z2	Annual investment in building forest parks	billions
z3	Area of new woodland created in the current year	1 hectare

.

The above data are also from the China Forestry and Grassland Statistical Yearbook for the years (2003, 2008, 2012, and 2018). Here, in order to simplify the content, we will not repeat the elaboration. After performing the benchmark regression, some tests are still needed for robustness testing, although we chose a two-way stationary model to avoid possible time-trendiness and individual characteristics. Here, we used quantile regression methods and outlier removal methods for robustness testing. Quantile regression can effectively demonstrate the extent to which different factors explain the dependent variable at different levels, which reflects the reliability of the results to some extent, while outlier removal is a common means in econometrics to minimize potential errors. Here, we chose to report the regression results at three quartiles: 50% and 75%. In the outlier removal test, we removed four municipalities.

3. Analysis of Results

Based on the above methodology, we measured carbon emissions from forest tourism and carbon sinks in forest parks in all provinces and cities in China. Figure 3 shows the results of carbon emissions and carbon sinks, with carbon emissions on the left and carbon sinks on the right.

According to Figure 3, China’s forest tourism carbon emissions and carbon sinks show significant differences in different regions. In the eastern region, Jiangsu, Guangdong, and Zhejiang have relatively high carbon emissions due to a more developed tourism industry, while Zhejiang and Fujian demonstrate strong carbon sinks, thanks to their high forest coverage and abundant forest park resources. Carbon emissions in the central region are generally low, with Henan, Hubei, and Hunan standing out, while these regions also demonstrate good carbon sinks, possibly related to their abundant forest resources. The western region has lower carbon emissions due to the lower level of tourism development and population density, but Sichuan, Yunnan, and Guizhou show excellent performance in carbon sinks due to their rich natural resources and vast forest areas. In terms of temporal trends, carbon emissions increased in most provinces with the expansion of tourism between 2003 and 2018, while carbon sink capacity increased in regions such as Zhejiang and Fujian. Of course, looking at the statistics only yields numerical patterns. For this reason, with the help of exploratory spatial analysis methods, we show the spatial characteristics of carbon emissions and sinks presented at different points in time, and Figure 4 and Figure 5 show the results for forest tourism carbon emissions and forest park carbon sinks, respectively.

According to Figure 4, the map from 2003 to 2018 shows that the spatial distribution of carbon emissions from forest tourism in China has changed significantly. Initially, carbon emissions were mainly concentrated in the economically developed provinces along the eastern coast, such as Jiangsu, Zhejiang, and Guangdong, which have high carbon emissions due to their well-developed tourism industry. Over time, the areas with high values of carbon emissions gradually expanded to the central and western parts of the country, especially Sichuan and Yunnan, reflecting the development of tourism resources and the increase in the number of tourists in these regions. By 2018, even carbon emissions in northeastern and remote western provinces, such as Xinjiang, had risen, showing that tourism had spread across the country, but overall showing the spatial characteristics of high in the south and low in the north. Overall, with the promotion and development of tourism, carbon emissions from forest tourism in China have not only increased, but also shifted from a pattern of concentration in the east to a wider distribution across the country. This spatial and temporal change demonstrates the interaction between tourism development and the growth of economic activity between regions, as well as the broader environmental impacts of tourism expansion.

According to Figure 5, the map from 2003 to 2018 shows that the spatial distribution of carbon sink capacity in China’s forest parks has experienced significant changes. During this period, the carbon sink capacity showed an overall trend of enhancement, especially in the central and western regions. Initially, regions with stronger carbon sink capacity were mainly concentrated in the southeastern coastal provinces such as Zhejiang, Fujian, and Guangdong, which are rich in forest resources. As time progresses, the carbon sink capacity of central and western provinces such as Sichuan, Yunnan, and Qinghai gradually increases, and dark green areas are becoming more and more common in these regions, reflecting the restoration and enhancement of forest ecosystems. This increase in carbon sink capacity from east to west reflects the general improvement in forest and ecological environmental protection efforts nationwide. In general, however, the areas with higher carbon sinks in forest parks are mainly in the northeast and southwest, and areas with scarce forest resources, like the plains of east and north China, have maintained their forest park carbon sinks at a lower level.

Next, we need to calculate the global Moran index for forest tourism carbon emissions and forest park carbon sinks to determine that they are indeed spatially correlated.

Table 3. Global Moran’s index.

Variable (Year)	Global Moran’s I	Z	p
FTEC2003	0.219	3.1291	0.005
FTEC2008	0.223	3.2310	0.005
FTEC2013	0.214	3.0715	0.006
FTEC2018	0.201	3.0226	0.006
CFP2003	0.337	3.0532	0.009
CFP2008	0.339	3.0689	0.008
CFP2013	0.344	3.0704	0.008
CFP2018	0.341	3.0667	0.008

reports the results of the Moran index test for FTEC and C_FP.

Table 3 reports the global Moran indices for FTEC and CFP, and the results show significant and positive spatial correlations over all years. However, there are not more significant fluctuations and variations, but this at least validates the feasibility of the spatial analysis. Further, we examined the spatial clustering characteristics of carbon sinks and carbon emissions through LISA clustering diagrams, which are illustrated in Figure 6 and Figure 7 for carbon emissions and carbon sinks, respectively.

In Figure 6, we find that in 2003, Gansu and Ningxia in the northwestern region showed high–low clustering, which indicates that for the entire western region, the forest tourism carbon emissions in these two regions were significantly higher than those in neighboring regions (Shaanxi, Qinghai, Inner Mongolia, etc.). On the other hand, the southeastern coastal region (including Shanghai, Anhui, Jiangxi, and Fujian) exhibits low–high clustering, which indicates that its forest tourism carbon emissions are significantly lower than those of neighboring regions (e.g., Jiangsu, Zhejiang, Hubei, and Hunan). At this stage, there was no obvious clustering of tourism carbon emissions. However, by 2008, high–low clustering disappeared in the northwestern part of the country and narrowed along the southeastern coast, suggesting that the rapid development of forest tourism may have broken past spatial imbalances, i.e., areas that had not yet been developed for forest tourism were strengthened at this stage. In 2013, a large area of low–low clustering was observed in the entire northern part of the country (Inner Mongolia, Hebei, Beijing, Gansu, Ningxia), which undoubtedly indicates that the development of tourism, especially forest tourism, is migrating from north to south, with Shanxi becoming the only region with a higher level. At this time, Hunan and Jiangxi, which are rich in forest resources, show high agglomeration. On the other hand, Guangxi, Fujian, and Hainan are the few low-carbon areas for forest tourism in the southern region. Finally, in 2018, the center of gravity of total forest tourism carbon emissions shifted to the south, with Hubei, Hunan, Jiangxi, and Guangdong showing high–high agglomeration, while Guangxi, Fujian and Hainan maintained low–high agglomeration.

Next, we analyzed the clustering of carbon sinks in forest parks. In 2003, the eastern region showed high–high aggregation, which is due to the very rich forestry resources in the northeast region. Correspondingly, almost all of the plains (as well as the southeastern hilly areas) showed low–low aggregation. This is due to the fact that forestry resources were not yet rationally developed and utilized at that time, and that these areas were the main production areas for food crops. This situation, in 2008, did not change significantly. It was not until 2013 that this pattern began to change, and the regions located in the south (Guangdong, Guangxi, Fujian, Zhejiang, Hunan) began to move away from this low–low aggregation. This was due to the fact that forest resources in the southern provinces were given sufficient attention, and the number and quality of forest parks were significantly improved. By 2018, only the plains of the middle and lower reaches of the Yangtze River and the plains of North China remained in low–low clustering, which is of course well explained because these are the main grain-producing areas of China, and of course do not have any outstanding forest resources.

Finally, we examined the relationship between carbon sinks in forest parks and carbon emissions from forest tourism seen Figure 8. It is clear that CFP far exceeds FTEC in most of the regions, with only Shandong, Zhejiang, Beijing, Shanghai, Tianjin, and Guizhou showing that FTEC exceeds CFP (with Shanghai consistently having more carbon emissions than carbon sinks). It can be seen that those areas with little forest resources but with high population density, strong consumption capacity, and a hot tourism market are prone to have carbon emissions exceeding carbon sinks. In particular, the cities of Beijing, Tianjin, and Shanghai are themselves relatively small.

So far, we have calculated and summarized the spatial distribution patterns and evolutionary processes of FTCE and CFP. However, for these conclusions, there are no clear numerical values, and so far they are just vague conclusions. In addition, the reasons for these results are only speculation rather than scientific analysis, so we used regression analysis methods to determine the regression analysis here. The first and third columns are the results without fixed effects (i.e., mixed regression), while the second column reports the results with fixed effects. The results are shown in

Table 4. Results of the analysis of influencing factors.

	(1)	(2)	(3)	(4)
	FTCE	FTCE	CFP	CFP
x1	0.595 ***	0.581 ***
	(4.64)	(4.53)
x2	0.433 ***	0.379 ***
	(4.64)	(4.64)
x3	0.799 ***	0.701 **
	(2.81)	(2.01)
z1			1.533 **	1.264 *
			(2.27)	(1.85)
z2			0.141	0.093
			(1.03)	(0.68)
z3			0.051	0.017
			(0.24)	(0.19)
Year	NO	YES	NO	YES
Province	NO	YES	NO	YES
N	120	120	120	120
Adj.R2	0.735	0.516	0.780	0.667

Note: *, **, and *** indicate 10%, 5%, and 1% significance levels, respectively, and values in parentheses are t-values.

.

As seen through the results for FTCE, total visitors, park revenue, and tourist highway miles exhibit positive impact effects with or without controlling for individuals and time of day. Specifically, the regression coefficients of x₁ are 0.595 and 0.581, respectively, and are significant at the 1% level, which indicates that the number of tourists is the main factor that increases forest tourism carbon emissions. Also, the regression coefficients of x₂ are 0.433 and 0.379, respectively, also significant at the 1% level, which indicates that, in addition to visiting forest parks, the spending of tourists in the process of traveling is also worthy of attention, and in general, the higher the expenses, the higher are the forest tourism carbon emissions. Finally, for x₃, which mainly reflects the degree of development of the forest park, the regression coefficients are 0.799 and 0.701, respectively, both remaining significant at the 5% level. This suggests that, in addition to tourists and revenue, the development of transportation roads in the scenic area itself is the main factor in expanding tourism carbon emissions, and its role is even more than that of tourist arrivals.

Next, we look at the differential performance of the influencing factors from a CFP perspective. In columns 3 and 4, we can find that neither z₂ nor z₃ pass the 10% test for significance. This suggests that investment in the development of forest parks and the addition of new forest land area are not the main drivers of the increase in forest parks; i.e., the additional forest land area appears to be very small compared to the increase in the number of forest parks, and at the same time the construction and development of forest parks does not increase the carbon sink of the forests, and it may increase the tourist pressure on the forest parks (as mentioned earlier). Only the number of employees in forest parks showed a significant positive effect, with regression coefficients of 1.533 and 1.246, respectively, which could be attributed to the fact that the employees are able to maintain the overall ecological environment of the forest parks to the extent that it is not destroyed, which plays a role in forest restoration.

In addition, apart from conducting global analysis, we also wanted to have a separate discussion on high-value clustering areas, such as high carbon sink and high carbon emission clusters. This is because the influencing factors in these regions may differ from the entire sample, especially in terms of the size and direction of the coefficients. Therefore, we considered the southern region as a heterogeneity test (including 11 provinces including Yunnan, Guizhou, Sichuan, Chongqing, Guangxi, Hubei, Hunan, Jiangxi, Fujian, Guangdong, and Zhejiang). In addition, to ensure the authenticity and credibility of our findings, we conducted robustness testing, including outlier removal testing and quantile regression. The results are shown in

Table 5. Robustness test and heterogeneity test.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	Southern region		Remove outliers		50% percentile		75% percentile
	FTCE	CFP	FTCE	CFP	FTCE	CFP	FTCE	CFP
x1	0.773 ***		0.483 ***		0.323 ***		0.581 ***
	(9.21)		(4.41)		(2.83)		(3.51)
x2	0.801 ***		0.321 ***		0.260 ***		0.550 ***
	(6.43)		(4.50)		(2.64)		(6.08)
x3	0.880 ***		0.797 **		0.399		0.520 *
	(4.05)		(2.06)		(1.43)		(1.65)
z1		1.026 **		1.315 **		0.641 **		0.795 *
		(2.27)		(1.99)		(2.21)		(1.90)
z2		0.402		0.088		0.080		0.087
		(0.91)		(0.71)		(0.46)		(0.52)
z3		0.009		0.015		0.003		0.008
		(0.29)		(0.17)		(0.09)		(0.12)
Year	YES	YES	YES	YES	YES	YES	YES	YES
Province	YES	YES	YES	YES	YES	YES	YES	YES
N	48	48	104	104	120	120	120	120
Adj.R2	0.712	0.704	0.691	0.643	0.417	0.405	0.557	0.503

Note: *, **, and *** indicate 10%, 5%, and 1% significance levels, respectively, and values in parentheses are t-values.

.

Here, the first two columns report the regression results of the southern region, namely the heterogeneity of influencing factors in a specific area. Southern China is the main distribution area of forest resources in China, as well as the gathering area with the largest number of forest parks. It is also an area where high-value carbon sinks are formed in forest parks. Therefore, it is necessary to discuss this region separately. We found that almost all influencing factors in this region, whether carbon emissions or carbon sinks, are greater and more significant, with all values greater than the national sample region. This indicates that regardless of the influencing factors, they play a greater role in areas where forest resources are already abundant. Similarly, the last six columns report the robustness results, and none of the coefficient directions have changed, indicating that our results are reliable.

4. Discussion

In this study, our contribution may be outlined as follows. First, in terms of research topics, we distinguished our research subjects from many previous studies. We did not investigate specific forest areas or forest parks, but instead we analyzed and discussed provincial forest parks nationwide based on publicly available data. Especially, the topic we discussed is relatively novel because we did not separately study the balance between forest carbon sinks and carbon emissions but further investigated the balance between forest tourism carbon emissions and forest parks from the perspective of forest tourism, which fully expands the value of forest parks in low-carbon tourism research. Particularly, we put forward our views on forest tourism, which is considered a key pathway for “ecotourism”. Second, in terms of supplementing existing literature, however, there are few studies that have investigated all forest parks at the provincial level. In addition, as an important component of ecotourism, forest tourism has always been a topic of wide concern. However, people rarely pay attention to the insufficient carbon emissions of forest tourism, but in fact, the carbon emissions of forest tourism should not be ignored. Our work proves that in areas with scarce forest resources, forest tourism may bring greater pressure on local carbon sinks (such as Beijing and Shanghai) [41,42]. Specifically, in 2018, the difference between forest tourism carbon emissions and forest park carbon sinks in Beijing was 3,529,907.492, while in Shanghai during the same period it was 2,038,269.021. These are two typical regions where carbon emissions exceed carbon sinks. However, the causes of these results have also been confirmed in the analysis of influencing factors, as tourism carbon emissions are positively affected by the number of tourists, tourism revenue, and tourism road mileage, respectively (their benchmark regression coefficients are 0.595, 0.433, and 0.799, respectively, and are significant at the 1% level). In fact, these data are also constantly growing. However, due to the limited availability of forest parks in Beijing and Shanghai, there has naturally been an overload situation. This is also the most important conclusion drawn in this article. We undoubtedly call for the reasonable protection and restriction of development of forest parks in areas with prosperous tourism but scarce forest resources.

In China, achieving the dual carbon goals of peak carbon emissions and carbon neutrality requires reducing carbon emissions and increasing carbon sinks. And forests have always been regarded by humans as the lungs of the earth on which survival depends. Forest tourism is indeed a low-carbon consumption method compared to traditional tourism, and this tourism industry undoubtedly promotes the peak and continuous decline of tourism carbon emissions. However, we cannot ignore the potential pressure it may impose on forest carbon sinks. As we just mentioned, in addition to typical areas such as Beijing and Shanghai, there are also some regions that need to pay attention to the pressure of tourism carbon emissions on local forest park carbon sinks. Because, in fact, it is often the people in areas with scarce forest resources who seek refuge in the forest. We can imagine that if you live near the forest, you will not go to the forest to play.

Finally, there are limitations in data selection. We have to recognize that all measurements and estimates are subject to error, and even though we have selected FGR data that appear to be the closest to the real situation, inevitably, temporal continuity is missing as a result. Specifically, we were unable to observe changes (which may or may not be linear) over a five-year period using this data set, and this is the greatest limitation of our data. However, we believe that methods to compensate for this limitation using other data sources may emerge in other similar studies or in some future studies. Moreover, there are some non-forestry lands such as water or grassland, rocks, etc. in the forest park, so the carbon sink value of the forest park cannot reflect the carbon sink situation of the forest park completely and accurately, which is only an overall estimation and speculation value, not the real value. However, we would like to pay more attention to the relationship between carbon sinks in forest parks and carbon emissions from forest tourism rather than accounting for them in a faulty way, so we are actively seeking a more appropriate measurement method in our future work.

5. Conclusions and Policy Implications

5.1. Conclusions

In this paper, we conducted a comprehensive assessment of carbon emissions and sinks in national forest parks from 2000 to 2019, revealing some key trends. First, the data estimated through the “top-down” method show that the carbon emissions of the southern tourism industry are usually higher than those of the northern region, mainly because the southern tourism industry is generally more developed. Numerically speaking, the average carbon emissions from forest tourism in the northern region in 2003 were 616,856.28, while the average in the southern region was 1,042,981.85. This situation where the south is greater than the north persisted until 2018.

However, in the eastern coastal areas, although the tourism industry is equally developed, the carbon emissions from forest tourism are not significantly different from those in the western regions due to relatively limited forest resources. This can be easily observed through the results of the East China and Southwest regions in 2018. The average of the six provinces in East China (Jiangsu, Anhui, Shandong, Fujian, Jiangxi, Zhejiang) is 12,368,968.23, while the average of the four provinces in Southwest China (Sichuan, Chongqing, Guizhou, Yunnan) is 15,608,963.22. This phenomenon highlights the significant impact of forest resource distribution on carbon emission patterns. In terms of carbon sequestration, forest parks in Northeast and Southwest China have been increasing their carbon sequestration, driven by both the increasing number of national forest parks and the fact that forest park site selection heavily relies on local natural resource endowments. Especially in eastern regions such as Jiangsu, Zhejiang, and Shanghai, due to the limited forest resources mainly composed of plains, farmland, and construction land, the carbon sink capacity of forest parks in Jiangsu, Zhejiang, and Shanghai has grown slowly. Specifically, the carbon sink of forest parks in Jiangsu, Zhejiang, and Shanghai is the lowest among all regions, increasing from 15,214,110.52 in 2003 to 62,034,578.22 in 2018, only by more than four times. With the passage of time and the strengthening of forest conservation policies, the carbon sequestration capacity in other regions such as Central China has significantly increased, indicating the importance of forest conservation and expansion in enhancing carbon sequestration.

Finally, we also analyzed potential influencing factors and found through heterogeneity and robustness tests that for forest tourism carbon emissions, the number of tourists, tourism revenue, and tourism road mileage all have a positive impact. However, for forest park carbon sinks, only the number of forest park employees can have a positive impact (regression coefficient of 1.553). This indicates that the development and protection of forest tourism may still be in an imbalanced state, and how to develop these forest parks reasonably while also protecting forest resources is an important topic.

5.2. Policy Implications

In this study, we conducted a comprehensive evaluation of the carbon emissions and carbon sink data of national forest parks from 2000 to 2019, and found significant differences in carbon emissions between different regions. Forest parks play an important role in achieving low-carbon tourism goals. This indicates that there is still a long way to go in protecting forests and ensuring their function as stable carbon sinks, and more policy support and technological innovation are needed to enhance the capacity and efficiency of forests as carbon sinks [43,44]. Based on the policy background of China’s current efforts to achieve peak carbon emissions and carbon neutrality goals, and our research findings, we propose the following policy recommendations outlined below.

First, in terms of management policies, it is recommended that the National Forestry and Grassland Administration work together with the tourism department to develop stricter regulations for the protection and utilization of forest parks. For areas with high tourism pressure, such as Beijing and Shanghai, more detailed visitor control policies should be implemented, including limiting the maximum daily number of tourists and introducing a tourist reservation system to ensure that tourism activities do not exceed the forest ecological carrying capacity. At the same time, intelligent monitoring systems need to be promoted, such as installing environmental sensors and real-time video surveillance, to ensure effective protection and management of forest resources through technological means. We call for reducing the impact of tourism activities on forests in areas where forest tourism is relatively scarce, by adopting protective policies for developing forest parks rather than expecting them to create more economic value.

Second, from the perspective of sustainable tourism development, investment in low-carbon tourism infrastructure should be increased, such as building charging stations around forest parks, encouraging the use of electric vehicles, and providing rental services within the park. In addition, low-carbon facilities such as solar lighting and recycled water systems can be considered for deployment within forest parks to reduce the carbon footprint of park operations. In terms of education and publicity, we will increase public awareness of environmental protection, and enhance tourists’ understanding of carbon emissions and environmental responsibility through exhibitions and information boards.

Once again, regarding policy recommendations on tourism carbon emissions, it is suggested that local governments establish specific carbon emission reduction targets and enforce the participation of the tourism industry in the carbon emission trading system through legislative means. For tourism projects with high carbon emissions, such as long-distance travel and large tour groups, carbon taxes can be levied to regulate them. At the same time, tourism enterprises are encouraged to invest in carbon offset projects, such as afforestation and renewable energy projects, to achieve carbon neutrality. In addition, it is recommended to carry out carbon footprint assessment services to provide tailored carbon reduction solutions for tourism enterprises and tourists.

Through these specific and detailed policy recommendations, the aim is to promote the rational use and protection of forest parks, while also driving the low-carbon transformation of the tourism industry, ensuring that economic benefits are pursued while also considering ecological protection, and contributing to China’s carbon neutrality goals [45]. The implementation of these policies will help China demonstrate leadership in global environmental governance and promote sustainable development and environmental protection both domestically and internationally.